Linear Control Systems Lecture # 23 Time-Varying Systems Solution of The State Equation

Linear Control Systems
Lecture # 23
Time-Varying Systems
Solution of the State Equation
p. 1/2
Consider the homogeneous equation

x = A(t)x,
x(t0 ) = x0
where the elements of the n n matrix A are continuous

functions of t over the period [t0 , t0 + T ]
x(t) is a solution of the differential equation if and only if it
satisfies the integral equation
x(t) = x0 +
A( )x( ) d
t0
p. 2/2
Successive Approximation
x0 (t) = x0
x1 (t) = x0 +
x2 (t) = x0 +
= x0 +
A(1 )x0 d1
t0
Z t
t0
Z t
A(1 )x1 (1 ) d1
A(1 )x0 d1
t0
..
.
A(1 )
t0
xk (t) = x0 +
A(2 )x0 d2 d1
t0
t
t0
A(1 )xk1 (1 ) d1
p. 3/2
Will xk (t) converge as k ?

xk (t) = x0 (t) + x1 (t) x0 (t) + + xk (t) xk1 (t)
= x0 +
k1
X
j=0
xj+1 (t) xj (t)
Z t
A(
)x
d
kx1 (t) x0 (t)k =
1
0
1
t0
Z t
kA(1 )k kx0 k d1 T kx0 k
t0
for all t [t0 , t0 + T ], where = maxt0 tt 0 +T kA(t)k
p. 4/2

Z t
kx2 (t) x1 (t)k =

A(
)[x
(
)
x
(
)]
d
1
1
1
0
1
1
t0
Z t
kA(1 )k kx1 (1 ) x0 (1 )k d1
t0
Z t
T kx0 k d1
t0
2
= T kx0 k(t t0 )
p. 5/2

Z t
kx3 (t) x2 (t)k =

A(
)[x
(
)
x
(
)]
d
1
2
1
1
1
1
t0
Z t
kA(1 )k kx2 (1 ) x1 (1 )k d1
t0
Z t
2 T kx0 k(1 t0 ) d1
t0
= T kx0 k
kxj+1 (t) xj (t)k T kx0 k
2 (t t0 )2
2!
j (t t0 )j
j!
T kx0 k
j T j
j!
p. 6/2
xk (t) = x0 +
k1
X
j=0
kxk (t)k kx0 k +
xj+1 (t) xj (t)
k1
X
kxj+1 (t) xj (t)k
j=0
kx0 k + T kx0 k
k1
X
j=0
k1
X
j=0
j T j
j!
j T j
j!
eT as k
lim xk (t) exists

k
p. 7/2
Thus, the infinite series

Z t
x(t) = x0 +
A(1 ) d1 x0
t0
A(1 )
t0
A(2 ) d2 d1 x0 +
t0
Is x(t) a solution? x(t0 ) = x0

x(t)
= 0 + A(t)x0 + A(t)
= A(t)[x0 +
= A(t)x(t)
A(2 ) d2 x0 +
t0
A(2 ) d2 x0 + ]
t0
p. 8/2
The solution x(t) can be expressed as
Z t
x(t) = I +
A(1 ) d1
t0
A(1 )
t0
A(2 ) d2 d1 + x0
t0
Define the n n matrix (t, t0 ) by

(t, t0 ) = I +
+
A(1 ) d1
t0
t
A(1 )
t0
A(2 ) d2 d1 +
t0
x(t) = (t, t0 )x0
p. 9/2
x(t) = (t, t0 )x0

(t, t0 ) is called the (state) transition matrix
The solution of x = A(t)x is unique (see Theorem 12.1,
page 54). Hence, (t, t0 ) is unique
p. 10/2
Properties of (t, ):
1.
(t, t) = I
2.
(t, ) = A(t)(t, )
3.
(t, ) = (t, )A( )
4.
(t, ) = (t, )(, ),
t, ,
p. 11/2
5. (t, )
is nonsingular for all t, and

1 (t, ) = (, t)
Proof of Property 3:
(t, ) = I +
+
A(1 ) d1
A(1 )
A(2 ) d2 d1 +
A(1 ) d1 = A( )
p. 12/2
A(1 )
= A( )
=
=
Z
Z
A(2 ) d2 d1
() d2 +
A(1 )
A(2 ) d2 d1
A(1 )A( ) d1
A(1 ) d1 A( )
(t, ) = I +
A(1 ) d1 + A( )
= (t, )A( )
p. 13/2
Proof of Property 4: Let x(t) be the solution of the state

equation x = A(t)x
x() = (, )x( )
x(t) = (t, )x() = (t, )(, )x( )
On the other hand,

x(t) = (t, )x( )
By the uniqueness of solutions

(t, )(, )x( ) = (t, )x( ),
x( )
(t, )(, ) = (t, )
p. 14/2
Proof of Property 5:
(t, )(, t) = (t, t) = I
This shows that (t, ) is nonsingular and

1 (t, ) = (, t)
p. 15/2
Complete Solution
Consider the state equation

x = A(t)x + B(t)u(t),
x(t0 ) = x0
where A(t) and B(t) are continuous functions of t and u(t)

is piecewise continuous function of t
The equation has a unique solution, given by
Z t
x(t) = (t, t0 )x0 +
(t, )B( )u( ) d
t0
Proof:
x(t0 ) = (t0 , t0 )x0 = x0
p. 16/2
x(t) = (t, t0 )x0 +
(t, )B( )u( ) d

t0
(t, t0 )x0 + (t, t)B(t)u(t)

t
Z t
+
(t, )B( )u( ) d
t0 t
= A(t)(t, t0 )x0 + B(t)u(t)
Z t
+
A(t)(t, )B( )u( ) d
x =
t0
= A(t) (t, t0 )x0 +
(t, )B( )u( ) d

t0
+ B(t)u(t) = A(t)x(t) + B(t)u(t)
p. 17/2
Special Case: A is constant

(t, ) = I +
+
A d1 +
Z
Z
A d2 d1
A d3 d2 d1 +
A d1 = (t )A
A d2 d1 =
(t )2
2!
A d3 d2 d1 =
(t, ) =
X
(t )k
k=0
k!
A2
(t )3
3!
A3
Ak = e(t )A
p. 18/2
Special Case: n = 1
x = a(t)x
dx
= a(t) dt
x
Z
ln
x(t)
x(t0 )
dz
z
x(t)
x(t0 )
x(t) = e
Rt
t0
a( ) d
t0
a( ) d
t0
a( ) d
(t, t0 ) = e
Rt
t0
x(t0 )
a( ) d
p. 19/2
Special Case: A is Diagonal
a1 (t)
A(t) =
(t, t0 ) =
a2 (t)
...
an (t)
1 (t, t0 )
2 (t, t0 )
...
n (t, t0 )
i (t, t0 ) = e
Rt
t0
ai ( ) d
p. 20/2
Special Case: A is Triangular
a11 (t) a12 (t) a13 (t)
A(t) =
a22 (t) a23 (t)
a33 (t)
11 (t, t0 ) 12 (t, t0 ) 13 (t, t0 )
(t, t0 ) =
22 (t, t0 ) 23 (t, t0 )
33 (t, t0 )
Solve the equations successively, starting from the third one

ii (t, t0 ) = e
Rt
t0
aii ( ) d
p. 21/2
Example
A=
"
2
t
0 1
x 1 = 2x1 + tx2 , x1 (t0 ) = x10

x 2 = x2 , x2 (t0 ) = x20
x2 (t) = e(tt 0 ) x20
Z t
x1 (t) = 11 (t, t0 )x10 +
11 (t, ) e( t 0 ) x20 d
t0
11 (t, t0 ) = e2(tt 0 )
p. 22/2
x1 (t) = e2(tt 0 ) x10 +

Z
"
e2(t ) e( t 0 ) x20 d
t0
e d = et (t 1) et0 (t0 1)
t0
x1 (t)
x2 (t)
"
e2(tt 0 ) 12 (t, t0 )
0
e(tt 0 )
#"
x10
x20
12 (t, t0 ) = e(tt 0 ) (t 1) e2(tt 0 ) (t0 1)

#
"
e2(tt 0 ) 12 (t, t0 )
(t, t0 ) =
0
e(tt 0 )
p. 23/2
Example:
x 1 = tx1 + tx2 ,
x2 (t) = e
x1 (t) = e
= e
2
2
1
(t
t
0)
2
2
2
1
(t
t
0)
2
Rt
t0
x1 (t0 ) +
x2 (t0 ) = e
Z
x1 (t0 ) + e
(t, t0 ) =
"
2
2
1
(t
t
0)
2
2
2
1
(t
)
2
x2 (t0 )
( 2 t20 )/2
x2 (t0 ) d
t0
2
2
1
(t
t
0)
2
1
(t2 t20 )
2
x 2 = tx2
1 2
2 (t
d x2 (t0 )
t0
1
(t2 t20 )
2
t20 )e
2
2
1
(t
t
0)
e2
p. 24/2

Linear Control Systems Lecture # 23 Time-Varying Systems Solution of The State Equation

Uploaded by

Copyright:

Available Formats

Linear Control Systems Lecture # 23 Time-Varying Systems Solution of The State Equation

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Linear Control Systems Lecture # 23 Time-Varying Systems Solution of The State Equation

Uploaded by

Copyright:

Available Formats

Linear Control Systems

Consider the homogeneous equation

where the elements of the n n matrix A are continuous

Will xk (t) converge as k ?

xj+1 (t) xj (t)

for all t [t0 , t0 + T ], where = maxt0 tt 0 +T kA(t)k

kx2 (t) x1 (t)k =

kx3 (t) x2 (t)k =

kxj+1 (t) xj (t)k T kx0 k

kxk (t)k kx0 k +

xj+1 (t) xj (t)

kxj+1 (t) xj (t)k

lim xk (t) exists

Thus, the infinite series

Is x(t) a solution? x(t0 ) = x0

The solution x(t) can be expressed as

Define the n n matrix (t, t0 ) by

x(t) = (t, t0 )x0

x(t) = (t, t0 )x0

(t, ) = (t, )A( )

(t, ) = (t, )(, ),

is nonsingular for all t, and

Proof of Property 4: Let x(t) be the solution of the state

On the other hand,

By the uniqueness of solutions

(t, )(, ) = (t, )

This shows that (t, ) is nonsingular and

Consider the state equation

where A(t) and B(t) are continuous functions of t and u(t)

x(t) = (t, t0 )x0 +

(t, )B( )u( ) d

(t, t0 )x0 + (t, t)B(t)u(t)

= A(t) (t, t0 )x0 +

(t, )B( )u( ) d

+ B(t)u(t) = A(t)x(t) + B(t)u(t)

Special Case: A is constant

Special Case: A is Diagonal

Special Case: A is Triangular

a11 (t) a12 (t) a13 (t)

11 (t, t0 ) 12 (t, t0 ) 13 (t, t0 )

Solve the equations successively, starting from the third one

x 1 = 2x1 + tx2 , x1 (t0 ) = x10

x1 (t) = e2(tt 0 ) x10 +

12 (t, t0 ) = e(tt 0 ) (t 1) e2(tt 0 ) (t0 1)

You might also like